1st International Conference on Industrial Networks and Intelligent Systems

Research Article

Work-In-Progress: Adaptive Population Artificial Bee Colony for Numerical Optimization

Download134 downloads
  • @INPROCEEDINGS{10.4108/icst.iniscom.2015.258339,
        author={Chun Ling Lin and Sheng Ta Hsieh and Shih Yuan Chiu},
        title={Work-In-Progress: Adaptive Population Artificial Bee Colony for Numerical Optimization},
        proceedings={1st International Conference on Industrial Networks and Intelligent Systems},
        publisher={ICST},
        proceedings_a={INISCOM},
        year={2015},
        month={4},
        keywords={adaptive population; artificial bee colony; cross-over; numerical optimization; population manager},
        doi={10.4108/icst.iniscom.2015.258339}
    }
    
  • Chun Ling Lin
    Sheng Ta Hsieh
    Shih Yuan Chiu
    Year: 2015
    Work-In-Progress: Adaptive Population Artificial Bee Colony for Numerical Optimization
    INISCOM
    ICST
    DOI: 10.4108/icst.iniscom.2015.258339
Chun Ling Lin1, Sheng Ta Hsieh2,*, Shih Yuan Chiu3
  • 1: Department of Electrical Engineering, Ming Chi University of Technology
  • 2: Department of Communication Engineering, Oriental Institute of Technology
  • 3: Systems Development Center, Chung-Shan Institute of Science and Technology
*Contact email: fo013@mail.oit.edu.tw

Abstract

In this paper, an adaptive population artificial bee colony (APABC) is proposed. In APABC, the population size of proposed ABC variant is not a fix but variable. The population size of APABC will be dynamically increased or decreased by population manager according to current solution searching status. It can enhance bees’ searching ability and increase population utilization. Thus, In order to test the efficiency of proposed method, fifteen test functions of CEC 2005, which include uni-modal and multi-modal functions, are adopted to test proposed method and compare it with original ABC. From the results, it can be observed that the APABC performs better on most test functions with 50 dimensions.